Goldman Sachs (GS) on Solana

Goldman Sachs is flagging a rapid deterioration in AI infrastructure debt as the six largest hyperscalers — Microsoft, Amazon, Alphabet, Meta, Oracle, Nvidia, and SpaceX — issued a combined $244 billion in bonds this year, more than double last year's total and 14 times 2024 levels. The bank reports that hyperscaler leverage ratios have nearly doubled, from 0.9x to 1.8x in roughly six months, as an estimated $5.8 trillion in combined AI capital expenditures through 2030 consumes most operating cash flow and forces heavy borrowing. Credit default swap spreads on major tech issuers have widened sharply, with the market's bond absorption capacity collapsing from $75 billion to just $25 billion.

The core risk Goldman identifies is a timing mismatch: companies are incurring massive upfront infrastructure costs that may not generate returns quickly enough to service the debt, potentially triggering rating pressure, higher financing costs, equity dilution, or forced cuts to capital spending. The bank signals that future valuations will shift toward accountability metrics — revenue growth, margin expansion, customer adoption, and free cash flow — rather than the scale of spending commitments alone.

Goldman Sachs raised its quarterly dividend by 25% following record Q2 2026 results, simultaneously announcing a $4 billion share repurchase program. The firm attributed the standout quarter to strength across investment banking, equities trading, and asset management, with management highlighting AI-related deal flow and a rebound in M&A activity as the primary growth drivers.

The combined capital-return package — a sizeable dividend increase paired with a fresh buyback authorization — signals management confidence in the durability of Goldman's earnings. Analysts note one watch item: dividend coverage by free cash flow, given the scale of the increase. GS shares have climbed roughly 26% year-to-date and more than 65% over the past year, reflecting investor conviction in capital-markets-focused banks as deal activity picks back up.

Goldman Sachs reported a blowout second quarter for 2026, with revenue rising 39.5% year over year to $20.34 billion, far exceeding the $16.13 billion analyst consensus. Earnings per share surged 92% to $20.98, nearly doubling the $14.48 estimate. All three core segments — investment banking, trading, and asset and wealth management — beat expectations, and the firm's efficiency ratio improved to 57.4%, a multiyear low. CEO David Solomon noted that despite the record results, the investment banking backlog grew to its highest level in five years and its second highest ever, driven by a record advisory backlog that reflects the breadth of client engagement.

The strong quarter prompted several analysts to raise their price targets on GS, with Wells Fargo lifting its target from $1,000 to $1,195 and Citigroup increasing its target from $930 to $1,100. CNBC's Investing Club also raised its own price target on the stock, citing the blowout quarter and an upbeat forward outlook underpinned by the firm's record deal pipeline.

Goldman Sachs is set to report Q2 2026 earnings on July 15, heading into the release with Wall Street analysts projecting sector-wide investment banking revenue to surge roughly 26% year-over-year, per KBW analyst Chris McGratty. The firm served as a lead underwriter on the SpaceX IPO — described as the largest IPO ever — alongside Morgan Stanley, earning an estimated $100 million in underwriting fees. Top-five banks collectively generated $11.1 billion in fees during the quarter, the highest tally in four-and-a-half years.

Geopolitical volatility from the Iran conflict provided an additional tailwind: sector trading revenue is forecast to climb 14% year-over-year as oil prices, interest rates, and currencies swung sharply. Beyond underwriting commissions, Goldman also stands to benefit from soft-dollar flows — fees from hedge funds seeking IPO share allocations — as well as debt financing mandates and wealth management work for SpaceX's newly minted employees. Goldman shares have risen more than 20% year-to-date heading into the print.

Goldman Sachs analyst Ronald Keung published a report on July 12 titled "Who Will Be the Long-Term Winner in China's AI Large Model Industry?" that identifies Zhipu and DeepSeek as the leading foundational text model players and ByteDance as the front-runner in multimodal and video generation. The framework evaluates competitors across pricing power, cost advantages, and financial strength, with MiniMax and Kuaishou also receiving Buy ratings. Zhipu, which received a Neutral rating with a $110 billion valuation, is noted for its open-source GLM model line, which Goldman views as broadly competitive with top US proprietary models.

A central investment thesis in the report is that Chinese AI models have achieved meaningful cost competitiveness, priced at roughly $1 per million tokens compared to $4–$8 for comparable US models — representing just 10–25% of US pricing. Goldman attributes this partly to smaller, more efficient architectures (2–10% of US model parameter counts) and Mixture-of-Experts routing techniques. The firm projects China's AI large model market revenue will grow roughly 25-fold from 35 billion RMB in 2026 to 879 billion RMB by 2030, driven by enterprise adoption and open-weight model releases — a trajectory Goldman believes positions select Chinese AI names as credible long-term winners in the global competition.

Goldman Sachs has staked out a major research conviction that optical networking — not compute — represents AI infrastructure's next dominant investment cycle. The firm's analysis argues that as data movement between processors becomes as critical as processing power itself, the addressable market for AI networking will expand from roughly $15 billion today to approximately $154 billion by 2028, a more than tenfold increase driven by optical technology displacing copper interconnects. Central to the thesis is co-packaged optics, which Goldman projects will capture around 59% of total AI networking spend by 2028, or roughly $91 billion in the base case. The research also highlights a dramatic increase in networking content value per compute unit: as Nvidia transitions from GB300 to Rubin Ultra architectures, networking spend per system is expected to rise nearly 29-fold, from about $315,000 to $9.4 million.

To anchor the call, Goldman identified Lumentum Holdings as what it considers the "purest" optical networking play, contrasting it with more diversified competitors such as Broadcom and Marvell Technology. The firm acknowledged the projection carries optimism and that co-packaged optics adoption could extend toward 2030 if implementation proves slower than anticipated. The report underscores Goldman's broader positioning as an active sector-conviction research house in the AI infrastructure build-out theme, following a string of recent calls on adjacent technology categories.

Goldman Sachs has secured outsourced chief investment officer (OCIO) mandates covering a combined $70 billion in retirement assets from Verizon Communications and Lockheed Martin, announced July 9, 2026. The mandates comprise roughly $30 billion in defined benefit pension assets across the two companies and $40 billion in Verizon defined contribution (401k) assets — one of the larger single OCIO wins in recent memory. Marc Nachmann, Goldman's global head of asset and wealth management, cited growing demand from large plan sponsors seeking a single partner with the investment depth to handle complex, bespoke retirement portfolios.

The wins expand Goldman's OCIO book to approximately $480 billion in assets, adding to prior mandates from UPS ($43.4 billion), Shell ($40 billion), and Eli Lilly ($25 billion). Goldman's broader asset and wealth management division now oversees roughly $3.7 trillion. The deals arrive as the US OCIO market — which grew 16% in 2025 and is projected to reach $4.3 trillion by 2030 — draws intensifying competition from BlackRock, Russell Investments, and Mercer, among others.

Goldman Sachs removed Duke Energy (DUK) from its US Conviction List on July 1, 2026, as part of a routine monthly update. The bank did not publicly cite specific reasons for the removal, but maintained its Buy rating on the utility stock, signaling continued positive sentiment despite the lower-conviction designation.

Duke Energy reported stronger-than-expected Q1 2026 results and has reaffirmed its full-year EPS guidance of $6.55–$6.80, alongside a 5–7% long-term growth target through 2030. The utility has also signed 7.6 GW of data center power agreements since 2024, a tailwind that has attracted broader analyst attention; Morgan Stanley raised its price target on the stock by $4 in late June, though it kept an Equal-Weight rating.

Goldman Sachs recently executed a series of fixed income offerings spanning maturities from 2027 to 2046, with fixed coupons and callable structures — a funding push that has drawn renewed scrutiny of the bank's valuation. A Simply Wall St. narrative DCF analysis now estimates GS shares at a fair value of $978.35, placing the current share price of $1,055.29 roughly 7.9% above that level.

The model applies measured revenue growth assumptions and profitability projections while benchmarking against a future P/E of 19.9x — well below the US Capital Markets industry median of 40.8x, where GS currently trades at approximately 19x. The analysis flags geopolitical disruptions to deal pipelines and rising regulatory compliance costs as the primary downside risks, while noting that GS's below-industry earnings multiple complicates a straightforward overvaluation read for investors weighing momentum against price.

Goldman Sachs economist Joseph Briggs is projecting that artificial intelligence could displace roughly 15 million American workers — about 9% of the U.S. workforce — over the next decade. The warning arrives against a soft labor market backdrop: June 2026 payrolls came in at just 57,000 new jobs, well below expectations, with Goldman estimating AI is already suppressing monthly payroll growth by 10,000–15,000 positions. Briggs anchors the 9% figure in a historical pattern where each 1% technology-driven productivity gain has raised job destruction rates by approximately 0.5 percentage points over two years; Goldman projects AI will ultimately deliver a 15% productivity boost at full adoption.

The strain is falling hardest on newer labor market entrants. Gen Z unemployment stands at 8.3%, roughly double the 4.2% national rate, while college graduates aged 22–27 face a 5.6% jobless rate. One-in-three employers surveyed by Goldman said they had replaced entry-level roles with AI rather than hiring. Customer service, back-office administration, and routine cognitive tasks are flagged as most immediately exposed, raising broader questions about whether productivity gains from AI will translate into wage growth or simply compress the bottom of the labor market.